Glutamine (GLN) is a classical nonessential amino acid, but becomes conditionally essential during catabolic states. In these conditions, relative GLN deficiency may contribute to diminished organ function, and underlie the limited impact of conventional nutritional support during severe illness. While both animal and clinical studies indicate benefit from GLN supplementation, physiocochemical properties of L-GLN limit provision in nutrient solutions. As an alternative, GLN dipeptides such as alanyl-GLN (ALA-GLN) confer both stability and solubility, and the dipeptides, are rapidly hydrolyzed to constituent amino acids in vivo. However, there have been no studies of GIN dipeptide supplementation in patients receiving enteral support - the preferred route of nutrient delivery - and there have been no comparisons of enteral vs. parenteral GLN dipeptide administration. To address these questions in a rigorous, comprehensive pilot clinical trial, our Specific Aims are: 1) To determine whether administration of ALA-GLN improves intestinal function, body protein, circulating immune cells, and systemic antioxidant capacity in critically ill patients requiring enteral tube feeding. 2) To determine whether intravenous is superior to enteral ALA-GLN administration. This will be a double-blind, randomized, controlled trial. Patients requiring tube feeding and intensive care will receive a standard, non- GLN enriched enteral formula, and isonitrogenous amino acids as: 1) iv ALA-GLN + enteral placebo; 2) iv GLN-free amino acids + enteral placebo; or 3) enteral ALA-GLN + iv placebo. Endpoints include: a) D-xylose absorption and gut permeability; b) nitrogen balance, 3-methylhistidine excretion, plasma GLN, and plasma insulin-like growth factor-I; c) total lymphocytes, T-lymphocytes, helper T-cells, and suppressor T-cells; and d) plasma glutathione (GSH) and GSH redox status, vitamins C and E, zinc, GSH peroxidase, and hydroperoxides. This study will provide important insight into the efficacy of novel methods of GLN administration in ICU patients, define possible mechanisms of GLN action, and generate important pilot data for subsequent multicenter trials.